Experimental Goodpasture’s syndrome in Wistar-Kyoto ratsimmunized with a3 chain of type IV collagen

MAURO ABBATE, RAGHURAM KALLURI,1 DANIELA CORNA, NAOTO YAMAGUCHI, ROBERT T. MCCLUSKEY,BILLY G. HUDSON, GIUSEPPE ANDRES, CARLA ZOJA, and GIUSEPPE REMUZZI

Mario Negri Institute for Pharmacological Research, Bergamo, Italy; Renal-Electrolyte and Hypertension Division, University ofPennsylvania Medical Center, Philadelphia, Pennsylvania; Pathology Research, Harvard Medical School, Massachusetts General Hospital-East, Charlestown, Massachusetts; Department of Biochemistry and Molecular Biology, University of Kansas Medical Center, KansasCity, Kansas; Department of Medical Sciences, Columbia Medical Center, New York, New York, USA; and Division of Nephrology andDialysis, Azienda Ospedaliera, Ospedali Riuniti di Bergamo, Bergamo, Italy

Experimental Goodpasture’s syndrome in Wistar-Kyoto rats im-munized with a3 chain of type IV collagen.

Background. Glomerulonephritis and lung hemorrhage of au-toimmune Goodpasture syndrome develop due to immune reac-tions against epitope(s) of the non-collagenous (NC1) domain ofa3-chain of type IV collagen [a3(IV) NC1]. Whether thymicmechanisms have a role in the loss of tolerance to the Goodpas-ture epitope has not been established. We studied the renal andpulmonary effects of immunization with different forms (mono-mer, dimer, or hexamer) of a3(IV) NC1 collagen in Wistar-Kyoto(WKY) rats, and assessed whether the intrathymic inoculation ofthe antigen may protect against anti-GBM disease.

Methods. WKY rats were immunized with bovine a3(IV) mono-mer, dimer, or hexamer, or with a3(IV) NC1 synthetic peptide.Renal function, kidney and lung immunohistology, and circulatingand tissue bound antibodies to type IV collagen chains wereanalyzed. Effects of intrathymic inoculation of antigen on subse-quent disease induction were analyzed in WKY rats given a3(IV)NC1 dimer or GBM preparation intrathymically 48 hours beforeimmunization.

Results. Proteinuria, linear IgG deposition in GBM, and cres-centic glomerulonephritis developed in WKY rats immunizedwith a3(IV) NC1 dimer or hexamer. Lesions were dose-depen-dent upon injections of 10 to 100 mg dimer. The a3(IV) NC1monomer induced less severe proteinuria and no crescents.Pulmonary hemorrhage was detectable in 35% of rats immunizedwith 25 to 100 mg a3(IV) NC1 dimer; a3(IV) synthetic peptide (36carboxyl terminal) did not induce disease. Rats injected intrathy-mically with up to 100 mg a3(IV) NC1 dimer or with GBM 48hours before immunization were not protected against subsequentdevelopment of proteinuria and glomerulonephritis.

Conclusions. These findings document that glomerulonephritisand lung hemorrhage can be elicited in WKY rats by immunization

with a3(IV) NC1. Failure of the intrathymic inoculation of antigen toprevent disease suggests that immunological tolerance cannot beachieved by this intervention, in contrast to other autoimmuneconditions, and may imply independent roles for cellular and hu-moral nephritogenic pathways in anti-GBM nephritis.

Some forms of human rapidly progressive glomerulone-phritis and hemorrhagic pneumonitis are induced by auto-antibodies to human basement membranes (Goodpasture’ssyndrome). During the last three decades several attemptshave been made to develop models that reproduce thehuman form of Goodpasture syndrome, to identify theantigen, and to understand the mechanisms of the autoim-mune response. At the beginning of 1960, Steblay showedthat sheep injected with crude preparations of either het-erologous or homologous glomerular basement membrane(GBM) develop antibodies to the injected antigen, andrapidly progressive glomerulonephritis. The antibodies bindto the GBM and incite the inflammatory reaction, as shownby their ability to transfer the disease to naive sheeps. Invitro the antibodies react with sheep alveolar basementmembrane, but deposition of immunoglobulin in the lungand hemorrhagic pneumonitis was not observed in ne-phritic sheep [1]. In 1984 Sado et al reported that Wistar-Kyoto (WKY) rats immunized with a glycoprotein purifiedfrom heterologous GBM develop rapidly progressive glo-merulonephritis and hemorrhagic pneumonitis [2, 3]. Therenal and pulmonary lesions were transferred to naive ratsby autoantibodies isolated from the urine of nephritic rats[4]. The development of autoimmune glomerulonephritisin WKY rats was confirmed by others [5, 6].

The antigen of human anti-GBM Goodpasture’s syn-drome is in the non-collagenous (NC1) globular domain oftype IV collagen [7, 8]. The relevant epitope(s) are presenton the alpha-3 chain of type IV collagen [9, 10]. It has beenreported that rabbits immunized with a3(IV)NC1 dimers–

1 Present address: Harvard Medical School, Beth Israel DeaconessMedical Center, Nephrology Division, Boston, MA 02215, USA.

Key words: extracellular matrix, dendritic cell, thymocyte, GBM a3(IV)NC1 peptide, lung hemorrhage, glomerulonephritis, anti-GBM disease.

Received for publication June 26, 1997and in revised form May 20, 1998Accepted for publication June 1, 1998

© 1998 by the International Society of Nephrology

Kidney International, Vol. 54 (1998), pp. 1550–1561

1550

but not with a1(IV)/a2(IV) NC1 dimers, or with the NC1hexamer [a1-a5(IV)NC1 domain]–develop autoantibodiesto a3(IV)NC1, glomerulonephritis and hemorrhagic pneu-monitis [11].

Autoimmune diseases occur when the immune system losesits capacity to distinguish between self and non-self [12]. Thethymus appears to play a major role in the development of selftolerance [13, 14], and also in tolerance to experimentalautoimmune disease [15–21]. Intrathymic injections of myelinbasic protein inhibit the development of experimental auto-immune encephalomyelitis (EAE) [15–17]. A single intrathy-mic injection of the major encephalitogenic peptide p71-90induces antigen-specific systemic tolerance, and inhibits theinduction of rat EAE [17]. Autoimmune diabetes in BB rats[18, 19] and in mice [20, 21] can also be prevented by thymicinjections of pancreatic islet cells. It is not known whether theintrathymic inoculation of the Goodpasture antigen may havesimilar protective effects against experimental autoimmuneanti-GBM disease.

We studied the autoimmune potential of heterologoustype IV collagen in WKY rats, using the monomer, thedimer and the hexamer forms of a3(IV) NC1 collagen asantigenic inoculum. Furthermore, we tested the effect ofintrathymic injections of these antigens on the developmentof experimental Goodpasture’s syndrome.

METHODS

Animals

Male WKY rats of 150 to 180 g initial body weight wereobtained from Charles River Italia s.p.a. (Calco, Italy) andused within 10 to 15 days for these studies (body weight atbaseline 200 to 230 g). Animal care and treatment wereconducted in conformity with the institutional guidelinesthat are in compliance with national (D. L. n. 116, G. U.,suppl 40, 18 Febbraio 1992, Circolare No 8, G. U., 14Luglio 1994) and international laws and policies (EECCouncil Directive 86/609, OJL 358, Dec 1987; Guide for theCare and Use of Laboratory Animals, U.S. National Re-search Council, 1996). All animals were housed in aconstant temperature room with a 12-hour dark/12-hourlight cycle and fed a standard diet.

Preparation of bovine GBM and type IV collagendomains

Glomerular basement membrane and a3(IV)NC1 hex-amers from bovine kidney were prepared as previouslydescribed [5, 10, 22, 23]. Glomeruli were isolated fromhomogenized cortical tissue using a differential sievingmethod. They were sonicated to obtain disrupted glomer-ular basement membrane suspension that was digested withcollagenase (Worthington CLSPA) at 37°C for 20 hourswith continuous stirring in 0.05 M Hepes buffer (pH 7.5)with 0.01 M CaCl2, and centrifuged at 100,000 3 g for onehour. Dimers and monomers of a3(IV)chain were pre-

pared from native bovine hexamer [24, 25] as previouslydescribed [10, 22]. Briefly, the NC1 hexamer was dissoci-ated, and the dimers were resolved from the monomers ona Sephacryl-200 column. Dimers were further resolved intofractions D1 and D2 on a C18 reversed phase HPLCcolumn, and monomers were likewise resolved into frac-tions M1, M2, and M3. Fraction D2 contained a3(IV)dimers (70%) and fraction M2 contained a3(IV) NC1monomers (100%) [10, 22, 26]. The 36-amino acid peptidein the C-terminal region was synthetically prepared by the“tea bag” method as described [10, 22]. The peptide wasnot coupled to KLH (keyhole limpet hemocyanine) or anycarrier protein. Briefly, the synthetic peptide was assembledby coupling on solid phase. A ninhydrin test was used todetermine the coupling efficiency after each step. Aftersuccessful coupling, the amino acid was deblocked for thenext round of synthesis. Upon completion of synthesis, thepeptide was cleaved from the resin and the functionalgroups deprotected using hydrofluoric acid. The 36-aminoacid peptide was purified using a C18 reversed phase HPLCcolumn and chemically identified by sequence analysis [26].

Immunization with bovine type IV collagen domains

WKY rats were immunized with the monomer, the dimerand the hexamer forms of the a3(IV)NC1 domains asantigenic inoculum according to either one of the followingprotocols:

Protocol 1. On day 0 the rats were administered theantigenic inoculum emulsified with equal volume of com-plete Freund’s adjuvant (CFA; Sigma Chemical Co., St.Louis, MO, USA) by foot pad injection, followed by asubcutaneous injection of 5 3 109 cells of pertussis vaccine(Michigan Public Health, Lansing, MI, USA). The ratswere sacrificed at days 28 to 35.

Protocol 2. On day 0, by foot pad injection the rats weregiven the antigenic inoculum emulsified with equal volumeof CFA, followed by subcutaneous injection of 5 3 109 cellsof pertussis vaccine. Fourteen days later the rats wereinjected again in different subcutaneous sites with the samedose of antigen as given on day 0, emulsified with equalvolume of incomplete Freund’s adjuvant (IFA; Sigma).Animals were sacrificed at day 35.

Experimental design I

To study the effect of immunization with bovine type IVcollagen peptides on kidney and lung pathology, ninegroups of rats were used (detailed in Table 1). Theexperimental groups included animals given bovine a3(IV)NC1 monomer (group IV), dimer (at increasing doses,groups V to VIII), and hexamer (group IX) as the antigenicinoculum accordingly to protocol 1 or protocol 2. Controlgroups were rats receiving either no treatment or phos-phate buffered saline (PBS) instead of the antigen at thetime of immunization.

Twenty-four hour urines were collected in metabolic cages

Abbate et al: Intrathymic antigen injection and nephritis 1551

on days 0 (baseline), 14, 21, 28 or 35 to measure urinaryprotein excretion. At the end of the study blood was collectedfor determination of serum creatinine and circulating antibod-ies to the a3 chain of type IV collagen. At sacrifice kidneysand lungs were removed and processed for immunofluores-cence and histological analysis and for detection of type IVcollagen a chain-specific antibodies by ELISA.

Experimental design II

To study the effect of intrathymic inoculation of GBMantigens on the development of experimental Goodpas-ture’s syndrome, seven additional groups of WKY rats wereused (Table 2). A series of experiments were designed in amodel of disease induced by a crude GBM preparation.The rats were injected intrathymically with 500 mg (250 mgin each lobe) of GBM emulsified in PBS (Group X), or withPBS alone (Group XI). After 48 hours the rats wereimmunized with 20 mg GBM emulsified with equal volumeof CFA by foot pad injection and subcutaneous injectionsat different sites. A subcutaneous injection with 5 3 109

cells of pertussis vaccine was given immediately afterwards.The effects of intrathymic injections of antigen were also

evaluated in rats in which the disease was elicited by usinga3(IV) NC1 dimer as the antigenic inoculum. The ratsreceived a3(IV) dimer at the dose of 25 mg (Group XII) or100 mg (Group XIII; 12.5 mg and 50 mg into each thymiclobe, respectively) dissolved in PBS. Control rats (group

XIV) were injected intrathymically with PBS alone. All ratswere immunized 48 hours later with 25 mg a3(IV) dimerwith the modalities reported in protocol 1.

Additional control experiments were performed to es-tablish the effect of intrathymic injection of synthetic a3(IV) chain peptide, which is not nephritogenic per se (asshown by data reported below), on subsequent develop-ment of disease. To this purpose rats were injected intra-thymically with 100 mg (50 mg in each lobe) a3(IV)synthetic peptide dissolved in PBS (Group XV) or withPBS (Group XVI). After 48 hours the rats were immunizedwith 25 mg a3(IV) dimer with the modalities reported inprotocol 1.

In three rats (Group XVII) the a3(IV) synthetic peptide(50 mg) was injected in the foot pad according to protocol1 in order to confirm the absence of nephritogenic potentialof the peptide preparation in our experimental conditions[10, 22].

Urinary protein excretion

Twenty-four hour urine samples were collected usingmetabolic cages and proteinuria determined by the modi-fied Coomassie blue G dye-binding assay for proteins withbovine serum albumin as the standard [27].

Elution of antibodies from tissues

After a flush and thorough washing of the organs withPBS to remove serum contaminants [11], kidneys and lungswere eluted with 0.025 M citrate buffer, pH 3.2, at roomtemperature. After neutralization and centrifugation, theeluates were concentrated by precipitation with 50% satu-rated ammonium sulphate and dialyzed against PBS. Fur-ther concentration was by negative pressure [28].

Enzyme linked immunosorbent assay (ELISA)

For direct ELISA, the plates were coated in triplicatewith 25 ng of antigen in 200 ml of coating buffer. Theantigen used to coat the plates was a mixture of bovinea3(NC1) dimers and monomers, purifed by HPLC [11, 29].Tissue eluates were analyzed by using recombinant humanNC1 monomers [30], and they revealed specific reactivitywith recombinant human a3NC1. The plates were incu-bated for two hours at 37°C or overnight at room temper-ature. Upon coating, the plates were washed three times atintervals of five minutes, with 0.15 M NaCl and 0.05%Tween 20 (washing buffer). After washing, the plates wereblocked with 2% BSA in PBS for 30 minutes at 37°C. Uponblocking, the plates were again washed with the washingbuffer and then incubated with primary antibody (ratcirculating and tissue bound antibodies, dilution 1:50) in anappropriate dilution in PBS. The plates were incubated forone hour at 37°C. After primary antibody incubation theplates were washed again and subsequently incubated withthe secondary antibody (goat anti-rat IgG) conjugated toalkaline phosphatase (AP) at 1:5000 dilution in PBS. The

Table 1. Experimental design I

GroupNo. of

ratsImmunization

(day 0)Immunization

(day 14)

I 3 None NoneII 3 PBS 1 CFA NoneIII 3 PBS 1 CFA PBS 1 IFAIV 2 monomer

10 mg 1 CFAmonomer 10 mg 1 IFA

V 5 dimer 10 mg 1 CFA NoneVI 4 dimer 25 mg 1 CFA NoneVII 5 dimer 25 mg 1 CFA dimer 25 mg 1 IFAVIII 5 dimer 50 mg 1 CFA dimer 50 mg 1 IFAIX 3 hexamer

400 mg 1 CFANone

Abbreviations are in the Appendix.

Table 2. Experimental design II

GroupNo. of

ratsIntrathymic injection

(day 22)Immunization

(day 0)

X 3 GBM 500 mg GBM 20 mgXI 4 PBS GBM 20 mgXII 3 dimer 25 mg dimer 25 mgXIII 4 dimer 100 mg dimer 25 mgXIV 4 PBS dimer 25 mgXV 2 recombinant a 3(IV)

peptide 100 mgdimer 25 mg

XVI 2 PBS dimer 25 mgXVII 3 — recombinant a 3(IV)

peptide 100 mg

Abbate et al: Intrathymic antigen injection and nephritis1552

plates were incubated for one hour at 37°C. Subsequently,the plates were again washed thoroughly, and substrate,disodium p-nitrophenyl phosphate (5 mg/ml) was added.After color development, the absorbance was measuredusing an ELISA autoreader at 405 nm [11, 24].

Histological methods

Samples of renal tissue were fixed in Dubosq-Brazil solu-tion, dehydrated in alcohol, and embedded in paraffin. Sec-tions (3 mm) were stained with Masson’s trichrome, hematox-ylin-eosin (H&E), and periodic acid-Schiff (PAS stain).Crescents were counted in at least 50 glomeruli randomlyselected in a histologic section from each animal. Crescentswere roughly divided into three groups: (a) circumferential;(b) those involving about one half of the glomerulus; and (c)those involving a quarter or less of the glomerulus [5]. Lungspecimens were fixed in 10% buffered formalin and processedas above. In some animals, lungs were inflated at sacrifice andfixed in the same fixative. The tissue sections were stained withH&E and by the Gomori’s Prussian Blue reaction for iron[31]. The slides were examined independently by two investi-gators in a blind fashion.

Immunofluorescence microscopy

Specimens of renal cortex and of lung were rapidlyfrozen in liquid nitrogen. The tissue fragments were em-bedded in OCT compound (Miles Laboratory, Ekhart, IL,USA) and sectioned (4 mm) using a Mikrom 500 O cryostat(Mikrom, Walldorf, Germany) for direct immunofluores-cence staining of rat IgG. Sections were fixed in acetone,blocked with 1% BSA in PBS, and incubated one hour atroom temperature with affinity-purified, fluorescein-conju-gated goat F(ab9)2 fragments (H1L) anti-rat IgG (JacksonImmunoresearch, West Grove, PA, USA). After washing inPBS, slides were mounted using 1% N-propyl gallate inTris-HCl 0.1 M, pH 8 to reduce fading, and examined withan Olympus BH-RFCA2 microscope equipped with epiflu-orescence using a USH-102D mercury lamp and appropri-ate optics. The evaluation was performed without knowl-edge of the group in at least 15 glomeruli from each rat ofeach group. The amount and extent of the immune depositswere semiquantitatively graded on a scale from 0 to 111

(0, absent; 1, minimal; 11, moderate; 111, marked andextensive) [5].

Electron microscopy

Small specimens of lung tissue were fixed in 2.5%glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) or withKarnovsky’s solution [32] for four hours at 4°C. Sampleswere washed in cacodylate buffer and subsequently post-fixed in 1% osmium tetroxide for one hour. After a briefwash in cacodylate buffer, they were dehydrated throughascending grades of alcohol and embedded in Epon resin.Sections were cut on an LKB V ultramicrotome. Semithinsections were stained with toluidine blue in borax andexamined by light microscopy. Ultrathin sections werestained with uranyl acetate and lead citrate and thenexamined with a Zeiss EM 109 electronmicroscope.

RESULTS

The main results are summarized in Tables 3, 4, 5, and 6;further details are given in the sections below. The gain inbody wt in the animals of the groups immunized withdifferent a3(IV)NC1 peptides was comparable to control(final body wts, mean 6 SD; control rats, groups I to III,307 6 24 g; rats immunized with a3 (IV) peptides, groupsIV to X, 292 6 27 g).

Renal and pulmonary effects of immunization withbovine a3(IV) NC1 monomer, dimer, and hexamer

Groups I to III. Normal rats and control rats injected withCFA only or CFA 1 IFA. The values of urinary proteinexcretion were within normal range. Kidneys and lungsshowed no histologic or immunohistochemical abnormali-ties.

Group IV. Rats immunized with a3(IV) NC1 monomer.These rats had increased urinary protein excretion rates(approximately 130 mg/day at sacrifice) and mild glomeru-lar histologic abnormalities mainly consisting of irregularhypercellularity of tufts and glomerulo-capsular adhesions.Linear GBM staining for IgG (111) was seen by immu-nofluorescence (Fig. 1A). Neither pulmonary lesions norlinear deposition of IgG in the alveolar basement mem-branes were detectable.

Table 3. Time course of urinary protein excretion (mg/day) in WKY rats immunized with a 3(IV) NC1 monomer, dimer or hexamer

Groups Day 0 Day 7 Day 14 Day 21 Final

I 20.6 6 2.2 17.8 6 2.2 22.9 6 5.2 28.4 6 5.0 23.4 6 4.8II 19.7 6 2.1 17.3 6 5.1 20.6 6 0.5 21.7 6 2.8 23.8 6 0.7III 15.8 6 4.5 20.5 6 4.2 16.2 6 3.8 16.5 6 3.3 29.7 6 24.6IVa 28 ND 42 68 134V 27.5 6 12.7 19.4 6 5.8 82.3 6 133.9 320.7 6 197.6 520.8 6 161.2VI 21.5 6 2.2 21.2 6 2.2 168.4 6 100.2 476.5 6 91.4 481.4 6 85.2VII 39.0 6 16.5 ND 88.5 6 40.0 229.0 6 21.7 385.7 6 213.8VIII 39.1 6 17.4 ND 100.0 6 14.5 162.0 6 74.0 283.0 6 141.5IX 20.1 6 3.3 19.6 6 2.4 115.1 6 114.8 315.9 6 231.5 399.2 6 250.8

Values are expressed as mean 6 SD. ND is not determined.a N 5 2.

Abbate et al: Intrathymic antigen injection and nephritis 1553

Groups V to VIII. Rats immunized with 10 to 100 mga3(IV) NC1 dimer: Proteinuria and renal lesions. These ratshad higher rates of urinary protein excretion as comparedwith both control animals and rats of group IV, withaverage values of 280 to 520 mg/24 hr at sacrifice. Glomer-ular damage was present in a dose-dependent fashion.Thus, most of the affected glomeruli in the rats injectedwith 10 mg a3(IV) NC1 dimer (group V) displayed eitherglomerulo-capsular adhesions or crescents involving only aquarter or less of their circumference, whereas almost no

glomeruli (0.2%) had circumferential crescents in thisgroup. The percentage of glomeruli with circumferentialcrescents progressively increased in the rats injected withtotal doses of 25, 50, and 100 mg (group VI, 8.0%; groupVII, 16.4%; and group VIII, 20%). The kidneys that werethe most severely damaged also contained fibrocellularcrescents with sclerosis of tufts, tubular casts, and irregulartubular atrophy. Immunofluorescence staining of kidneysections for rat IgG (Fig. 1B) revealed linear deposits alongthe GBM (111).

Table 4. Time course of urinary protein excretion (mg/day) in WKY rats injected intrathymically with crude GBM, a 3(IV) NC1 dimer orrecombinant a 3(IV) peptide

Groups Day 0 Day 7 Day 14 Day 21 Final

X 13.9 6 3.4 17.0 6 6.4 39.2 6 49.1 76.3 6 8.1 236.7 6 179.6XI 16.9 6 4.8 20.6 6 5.2 55.8 6 47.0 191.3 6 57.2 149.9 6 32.7XII 17.8 6 5.3 18.6 6 9.1 36.9 6 24.1 327.0 6 143.7 525.7 6 370.8XIII 19.5 6 14.7 18.6 6 2.6 29.2 6 37.6 302.2 6 190.4 270.4 6 183.9XIV 24.8 6 23.4 17.8 6 5.6 21.8 6 18.8 341.9 6 154.2 386.0 6 208.7XVa 19 ND 77 233 300XVIa 15 ND 25 212 449XVII 20.2 6 3.0 22.0 6 2.0 19.8 6 2.1 22.4 6 5.9 16.4 6 2.9

Values are expressed as mean 6 SD. ND is not determined.a N 5 2.

Table 5. Renal and lung pathology in WKY rats immunized with a 3(IV) NC1 monomer, dimer or hexamer

Groups

I(N 5 3)

II(N 5 3)

III(N 5 3)

IV(N 5 2)

V(N 5 5)

VI(N 5 4)

VII(N 5 5)

VIII(N 5 5)

IX(N 5 3)

KidneyGlomerular crescentsCircumferential % 0 0 0 0 0.2 6 0.2 8.0 6 4.6 16.4 6 1.8 20.0 6 7.0 01/2 Glomerulus % 0 0 0 0 18.6 6 6.7 31.2 6 7.6 25.2 6 4.0 22.0 6 5.8 19.0 6 7.61/4 Glomerulus % 0 0 0 0 42.0 6 9.0 46.0 6 6.7 44.4 6 2.6 16.0 6 4.0 69.4 6 8.7Rat IgG in GBMa 0 0 0 111 111 111 111 111 111

LungHemorragic lesions (%)b 0 0 0 0 0 50 20 40 0Rat IgG in ACWc 0 0 0 0 0 0 0 20 0

Values are expressed as mean 6 SE. Abbreviations are: GBM, glomerular basement membrane; ACW, alveolar capillary wall; ND, not determined.a Intensity of linear staining, score 0 to 111b Percentage of rats with lesionsc Percentage of rats with IF positive staining.

Table 6. Renal and lung pathology in WKY rats immunized with crude GBM, a 3(IV) NC1 dimer or recombinant a 3(IV) peptide

Groups

X(N 5 3)

XI(N 5 4)

XII(N 5 3)

XIII(N 5 4)

XIV(N 5 4)

XV(N 5 2)

XVI(N 5 2)

XVII(N 5 3)

Kidney, glomerular crescentsCircumferential % 0 0 0 0.2 6 0.2 1.0 6 0.4 0 1.1 01/2 Glomerulus % 8.7 6 2.9 14.5 6 5.4 17.0 6 2.3 19.2 6 3.9 19.7 6 1.2 7.8 20.5 01/4 Glomerulus % 55.7 6 10.9 63.5 6 3.4 57.6 6 7.7 41.2 6 1.9 53.5 6 6.5 66.8 62.3 0Rat IgG in GBMa 111 111 111 111 111 111 111 0

Lung Hemorragic lesions (%)b ND ND 0 0 0 0 0 0Rat IgG in ACWc ND ND 0 0 0 0 0 0

Values are expressed as mean 6 SE. Abbreviations are: GBM, glomerular basement membrane; ACW, alveolar capillary wall; ND, not determined.a Intensity of linear staining, score 0 to 111b Percentage of rats with lesionsc Percentage of rats with IF positive staining.

Abbate et al: Intrathymic antigen injection and nephritis1554

Groups V to VIII. Rats immunized with 10 to 100 mga3(IV) NC1 dimer: Lung pathology. The histological analy-sis of lungs from rats injected with 10 mg a3(IV) NC1 dimer(group V) showed no differences compared to control rats.In contrast, small hemorrhagic areas were detectable inlungs from 50% of rats immunized with 25 mg a3(IV) NC1dimer (group VI; Fig. 2B). Foci of hemorrhagic pneumo-nitis, characterized by accumulation of erythrocytes andmacrophages containing crystals of hemosiderin, were alsopresent in the lungs of rats injected with 50 mg (group VII;Fig. 2C) or 100 mg a3(IV) NC1 dimer (group VIII; Fig.2D); the percentage of rats with hemorrage was 20% and40%, respectively. The hemorrhagic lesions were moreevident in these groups as compared with rats immunizedwith 25 mg dimer, and they were severe in one of the

animals injected with 100 mg dimer. As shown in Figure 2D,the presence of hemosiderin in the crystals was confirmedby positive reaction with Prussian blue staining. Electrondense structures corresponding to the hemosiderin crystalswere demonstrated by electron microscopy in the cyto-plasm of macrophages and in the alveolar spaces withinhemorragic lesions (Fig. 3). Overall, 35% of the animalsdeveloped pulmonary hemorragic lesions after immuniza-tion with 25 to 100 mg a3(IV) NC1 dimer (groups VI toVIII) in this study. Modest infiltration of mononuclear cellsand thickening of alveolar septa were observed in lungs thatwere inflated and artificially distended at sacrifice.

Linear deposits of IgG were not generally detectable inlungs by immunofluorescence, including most of the lungswith hemorrhagic lesions. However, linear staining of rat

Fig. 1. Photomicrographs of cryostat sections of kidneys taken from rats immunized with 20 mg a3(IV) NC1 monomer (A), 25 mg a3(IV) NC1 dimer(B), or 400 mg a3(IV) NC1 hexamer (C), and from a control rat (D). Sections were stained for rat IgG by direct immunofluorescence. Rat IgG isdistributed in linear pattern along the glomerular basement membranes in the glomeruli from rats injected with the bovine a3(IV) NC1 peptides (A,B, C). No linear deposits of rat IgG are present in control kidney (D) (3500).

Abbate et al: Intrathymic antigen injection and nephritis 1555

Fig. 2. Pulmonary effects of immunization with bovine a3(IV) NC1 dimer in rats. (A) Normal lung structure in a control rat (H&E, 3150). (B) Sectionof lung from a rat immunized with 20 mg a3(IV) NC1 dimer exhibiting foci of hemorrhagic pneumonitis and mild thickening of alveolar septa (PAS;3150. (C) Rat immunized with 50 mg a3(IV) NC1 dimer. Red blood cells and macrophages containing hemosiderin are present in alveolar spaces(H&E, 3300). (D) Red blood cells and iron-containing macrophages are present in the alveolar spaces in the lung from a rat immunized with 100 mga3(IV) NC1 dimer. The inset shows macrophages containing crystals of hemosiderin (Gomori’s Prussian Blue; 3300; inset 3800).

Abbate et al: Intrathymic antigen injection and nephritis1556

IgG in alveolar capillary walls was present in focal areas inthe rat with severe hemorrhagic pneumonitis injected with100 mg dimer (Fig. 4).

Circulating and tissue-bound antibodies to a3 chain oftype IV collagen

The sera and tissues of rats injected with 25 mg a3(IV)NC1 dimer were analyzed for the detection of antibodies tothe a3 chain of type IV collagen by ELISA. Antibodies tothe a3 chain of type IV collagen were detected both inserum and lungs (adsorbance at 405 nm, mean 6 SE; serum,1.63 6 0.14; lung, 0.38 6 0.02; . fivefold increase com-pared to control values in groups I to III, 0.31 6 0.11 and0.07 6 0.01, respectively), and the highest levels were foundin the kidney eluates (1.13 6 0.14, . tenfold increasecompared to control, 0.10 6 0.02).

Specificity of kidney-bound antibodies to a3(IV) NC1domain

Kidney eluates were analyzed by ELISA for the presenceof antibodies against other type IV collagen NC1 domains(a1 to a6). As summarized in Figure 5, kidney-boundantibodies in rats immunized with 25 or 50 mg a3(IV) NC1dimer bound strongly to the a3(IV) NC1 recombinanthuman antigen and had minimal reactivity to other type IVcollagen NC1 domains.

Group IX rats immunized with a3(IV) NC1 hexamer. Therats had proteinuria, prevalent crescents of small size, andIgG deposition in their GBMs. None of these animals hadhemorragic pneumonitis or linear IgG deposits in lungs.

Effects of intrathymic injection of bovine GBM or a3(IV)NC1 domain on the development of anti-GBM disease

Groups X to XIV. A preliminary set of experiments wasperformed to establish whether the intrathymic injection ofantigen may protect WKY rats against anti-GBM glomer-ulonephritis in a model of disease induced by immunizationwith bovine GBM. As with the a3(IV) NC1 peptides, thedisease was characterized by proteinuria and formation ofcrescents of small size affecting a high percentage ofglomeruli (Table 6, group XI). Cellular and fibrocellularcrescents were associated with increase of mesangial matrixor partial sclerosis of tufts, and tubular cast formation. Thepretreatment of rats with intrathymic injection of GBMpreparation (group X) did not inhibit proteinuria and renallesions in this model.

Similar results were obtained in animals in which anti-GBMdisease was elicited by immunization with 25 mg bovinea3(IV) NC1 dimer. There was no reduction in the values ofproteinuria and numbers of crescents in animals that receivedup to 100 mg a3(IV) NC1 dimer into the thymus. Theformation and linear binding of rat IgG to GBM as detectedby immunofluorescence staining was not inhibited. Effects onlung disease were not established because none of the animalsdeveloped lung injury in these experiments.

Groups XV to XVII: a3(IV) synthetic peptide. Neitherproteinuria (16.4 6 1.7 mg/day at sacrifice) or signs of renaland lung injury were detected in animals immunized witha3(IV) synthetic peptide (group XVII). The rats givena3(IV) synthetic peptide into the thymus and then immu-nized with bovine a3(IV) NC1 dimer (group XV), devel-oped proteinuria and glomerular lesions of comparable

Fig. 3. Rat immunized with 50 mg a3(IV) NC1dimer. Electron micrograph of an hemorrhagicarea of lung showing red cells, in alveolarspaces, electron dense crystals of hemosiderin(arrows), and inflammatory cells (asterisks)(uranyl acetate and lead citrate; 33,300).

Abbate et al: Intrathymic antigen injection and nephritis 1557

severity to the rats given intrathymically PBS instead ofa3(IV) synthetic peptide (group XVI).

DISCUSSION

Very few experimental models have been described sofar for human Goodpasture’s syndrome demonstrating thatactive immunization with basement membrane antigens iscapable of inducing in association the typical renal and

pulmonary lesions [2, 10, 33]. A purified preparation ofa3(IV) NC1 dimer did so in rabbits by eliciting theformation and binding of pathogenic antibodies to theGBM and to the alveolar basement membranes [10]. Thedevelopment of rat autoimmune models using purifiedantigen is of major relevance in the light of recent thera-peutic approaches to anti-GBM glomerulonephritis in rats[3, 5], which may also attenuate lung hemorrhage, and as a

Fig. 4. Immunofluorescence micrographs of sections of lungs taken from rats injected with a3(IV)NC1 dimer or vehicle, stained for rat IgG. No lineardeposits of IgG are seen in the lungs of rats injected with 25 mg a3(IV)NC1 dimer (A) or with vehicle (B). (C) Linear deposits of IgG focally in alveolarcapillary walls of a rat immunized with the highest dose (100 mg) of a3(IV)NC1 dimer (3600).

Abbate et al: Intrathymic antigen injection and nephritis1558

model to study antigen-specific forms of prevention andtherapy. We found that the immunization of WKY rats withsoluble a3(IV) peptides induced glomerulonephritis andhemorrhagic pneumonitis resembling human Goodpas-ture’s disease. Proteinuria and glomerular extracapillarylesions were elicited in the animals immunized with a3(IV)NC1 dimer as antigenic inoculum. The immunization witha3(IV) NC1 hexamer also caused proteinuria and glomer-ular lesions in WKY rats, at least with the relatively highdoses used in this study, at variance with the lack of effectsseen in New Zealand White rabbits injected with smalleramount of hexamer [11]. As in human Goodpasture’ssyndrome and in the rabbit model [11], glomerular damageinduced by the a3(IV) chain in WKY rats was associatedwith the formation of circulating anti-a3(IV) NC1 antibod-ies and linear deposition of rat IgG in glomerular basementmembranes, reflecting the ability of a3(IV) chain antigenpurified from bovine GBM to elicit the production ofanti-GBM antibodies. Findings that antibodies in kidneyeluates were directed against a3(IV) NC1, with minimalreactivity to other type IV collagen chains, indicated thatonly the a3(IV) NC1 domain contained the targetepitope(s) for nephritogenic antibodies in this model. Incontrast to soluble a3(IV) dimer and hexamer purifiedfrom GBM, our data with a3(IV) synthetic peptide incontrol experiments confirm previous findings that theactive immunization of WKY rats with synthesized peptidecontaining the a3(IV) NC1 epitope did not cause disease,even despite the formation of anti-peptide epitope antibod-ies [34]. In this respect, however, a form of anti-GBM

nephritis has been induced most recently in WKY rats byinjection of a3(IV) NC1 domain prepared by recombinantDNA technology using mammalian cells, that may maintainthe pathogenic behavior of the native peptide [35].

Besides inflammatory changes in glomeruli, focal hem-orrhagic lesions were detected in lungs of WKY ratsimmunized with 25, 50, or 100 mg dimer. No effects on thelung were seen in rats immunized with 10 mg a3(IV) dimer.Pulmonary lesions were confined to small foci of hemor-rage in rats immunized with the dose of 25 mg, andcontained hemosiderin-laden macrophages in rats immu-nized with 50 or 100 mg dimer. Only the high dose ofa3(IV)dimer (100 mg) induced diffuse and severe hemor-ragic pneumonitis in one animal. Thus, like in kidney, thedevelopment of severe injury in lung of WKY rats appearsto be dependent on the amount of injected antigen. Lineardeposits of IgG in alveolar capillary walls were only occa-sionally detectable by immunofluorescence staining oflungs, this despite the presence of circulating antibodies tothe a3(IV) NC1 domain. In previous studies no linear IgGdeposits were found in the lungs of sheep or rats withanti-GBM glomerulonephritis [2, 3, 31]. The absence ofIgG deposits was attributed to the presence of the barrierprovided by the not fenestrated alveolar endothelium, thatlimits the access of circulating antibodies [36–38]. Studieshave actually shown that the lungs of animals injected withnephritogenic doses of anti-basement membrane IgG didnot exhibit IgG deposits and hemorrhage unless the per-meability of the alveolar endothelium was artificially in-creased by exposure to high concentrations of oxygen [37,38]. Beside viral infections, volatile hydrocarbons andsmoking may cause alveolar wall injury and in so doingincrease the susceptibility to pulmonary hemorragic mani-festations [39, 40]. Deposits of heterologous anti-basementmembrane IgG, focal hemorragic lesions and leukocyteaccumulation were also induced in mice by increasing thepermeability of the alveolar capillary wall with injections ofinterleukin-2 (IL-2) and interferon-a (IFN-a), polypeptidemediators generated by activated lymphoid cells and bycells infected by viruses, thus mimicking human Goodpas-ture’s disease [41]. Similar pathways of injury may partici-pate in the development of polmonary lesions in WKY ratsimmunized with a3(IV) collagen, particularly cytokineselaborated by lymphocytes and macrophages that infiltratethe lung, and possibly released in blood from the inflamedkidney. However, the precise mechanisms underlying lungdamage in the absence of detectable alveolar basementmembrane deposits of IgG in rats immunized with a3(IV)dimer remain uncertain, and might involve the participa-tion of specific T-cells that infiltrate the lung duringantibody-independent immune reactions.

One of the implications in the failure of intrathymicinoculation of antigen against anti-GBM glomerulonephri-tis from our data are that the animals did not achieve astate of immunological tolerance, at least not in pathways

Fig. 5. Specificity of kidney-bound antibodies from rats immunized witha3(IV) NC1 dimer (groups VI and VII) or hexamer (group IX) andcontrol (group III). Samples were assayed in triplicate for each rat. Barsrepresent three rats per group. Symbols are: (M) control; (f) 25 mg dimer;(u) 50 mg dimer; (M) hexamer. Values are mean 6 SE.

Abbate et al: Intrathymic antigen injection and nephritis 1559

critical to tissue injury unlike models of diabetes [18–21],autoimmune encephalomyelitis [15–17], and allograft rejec-tion [42] in which such intervention was effective. Evidencefor cell-mediated mechanisms in anti-GBM disease hasemerged from recent studies in rats [6] and mice [43].However, such nonhumoral mechanisms have role in theexpression of anti-GBM injury in some MHC haplotypesonly, and they seem to be engaged by passive transfer ofsyngeneic anti-a3(IV) NC1 antibodies to induce disease[43]. Whether this applies to the WKY rat is not known,nor whether to be effective the intrathymic challenge toantigen should interfere both with humoral and nonhu-moral pathways. A dissociation of effects was shown inexperimental autoimmune myasthenia gravis, induced inrats by immunization with acetylcholine receptor, in whichthe intrathymic inoculation of antigen had no effect on theproduction of anti-acethylcholine receptor antibodies whileinhibiting the proliferative response to the antigen bylymph node cells [44]. In our model, Goodpasture antigenwas injected intrathymically. In this setting, we speculatethat, due to the possible cryptic nature of Goodpastureepitopes [8, 19], they may not be processed and presentedby thymic epithelium [45] or by macrophage/dendritic cells[46]. It is also possible that the relevant CD41/CD81 cellpopulations have relatively low-affinity T cell receptors [47,48]. These factors allow potential autoreactive cells tosurvive deletion. Alternatively, dose-related factors mayhave importance. Studies using soluble alloantigen in renal[42] and cardiac [49] transplantation have shown that theoptimal dose of antigen is an important factor for thetolerogenic efficacy of intrathymic inoculation. This sug-gests that higher doses of a3(IV) NC1 than available in thisstudy were required to interfere with the intrathymicprocesses that have roles in the production of pathogenicantibodies and/or autoreactive T cells. It is interesting thatB cells [50–52], as well as CD41Th2 cells providing helpfor B cell antibody response in mice [53], are particularlyresistant to tolerance induction so as to require abundantamounts of antigen in studies of antigen feeding [54–56].

In summary, experimental Goodpasture’s syndrome wasinduced in WKY rats by immunization with purified prep-arations of soluble a3(IV) NC1 peptides. Disease wascharacterized by circulating anti-a3(IV) NC1 antibodies,proteinuria, crescentic glomerulonephritis with linear de-posits of IgG along glomerular basement membranes, andfoci of lung hemorrhage. Kidney-bound antibodies weredirected against the a3(IV) NC1 domain with minimalreactivity to the other type IV collagen chains. Glomerularlesions were dose-dependent, and severe lung hemorrhagewith linear IgG deposits in alveolar capillary walls devel-oped after injection of 100 mg a3(IV) NC1 dimer. Intra-thymic inoculation of GBM or soluble a3(IV) NC1 dimer48 hours before immunization did not prevent disruption ofself-tolerance by active immunization with pathogenic an-tigen. Further approaches using larger amounts or alterna-

tive forms of antigen should be made to clarify the role ofintrathymic events in experimental Goodpasture’s syndromeand to design antigen-specific therapies, possibly combinedwith depletion of peripheral autoreactive cells (anti-lympho-cyte serum) or with strategies to block T cell-B cell interac-tions and consequent formation of pathogenic antibodies.

ACKNOWLEDGMENTS

This work was supported by NIH Grants DK-36807 (G. Andres, PI) andDK-18381 (B. Hudson, PI). The expert technical assistance of Mr. ParvinTodd is appreciated. The authors are indebted to Ms. Laura Arioli forhelping preparing the manuscript and to Mr. Matteo Capitanio forphotographic work.

Reprint requests to Dr. Mauro Abbate, “Mario Negri” Institute for Phar-macological Research, Via Gavazzeni 11, 24125 Bergamo, Italy.E-mail: abbate@irfmn.mnegri.it

APPENDIX

Abbreviations used in this article are: AP, alkaline phosphatase; BSA,bovine serum albumin; CFA, complete Freund’s adjuvant; EAE, experi-mental autoimmune encephalomyelitis; ELISA, enzyme-linked immu-nosorbent assay; GBM, glomerular basement membrane; H&E, hematox-ylin-eosin stain; HPLC, high-pressure liquid chromatography; IFA,incomplete Freund’s adjuvant; KLH, keyhole limpet hemocyanine; NC1,non-collagenous domain; PAS, periodic acid Schiff stain; PBS, phosphatebuffered saline; WKY, Wistar-Kyoto rat.

REFERENCES1. STEBLAY RW: Anti-glomeular basement membrane glomerulonephri-

tis. Am J Pathol 96:875–878, 19792. SADO Y, OKIGAKI T, TAKAMIYA H, SENO S: Experimental autoimmune

glomerulonephritis with pulmonary hemorrhage in rats. The dose-effectrelationship of the nephritogenic antigen from bovine glomerular base-ment membrane. J Clin Lab Immunol 15:199–204, 1984

3. SADO Y, NAITO I, AKITA M, OKIGAKI T: Strain specific responses ofinbred rats on the severity of experimental autoimmune glomerulo-nephritis. J Clin Lab Immunol 19:193–199, 1986

4. SADO Y, NAITO I, OKIGAKI T: Transfer of anti-glomerular basementmembrane antibody-induced glomerulonephritis in inbred rats withisologous antibodies from the urine of nephritic rats. J Pathol 158:325–332, 1989

5. NISHIKAWA K, GUO Y-J, MIYASAKA M, TAMATANI T, COLLINS BA, SYM-S, MCCLUSKEY RT, ANDRES G: Antibodies to intercellular adhe-sion molecule 1/lymphocyte function-associated antigen 1 preventcrescent formation in rat autoimmune glomerulonephritis. J Exp Med177:667–677, 1993

6. BOLTON KW, MAY WJ, STURGILL BC: Proliferative autoimmuneglomerulonephritis in rats: A model for autoimmune glomerulone-phritis in humans. Kidney Int 44:294–306, 1993

7. WIESLANDER J, BARR JF, BUTKOWSKI RJ, EDWARDS SJ, BYGREN P,HEINEGARD D, HUDSON BG: Goodpasture antigen of the glomerularbasement membrane: Localization to noncollagenous regions of typeIV collagen. Proc Natl Acad Sci USA 81:3838–3842, 1984

8. WIESLANDER J, LANGEVELD J, BUTKOWSKI R, JODLOWSKI M, NOELKENM, HUDSON BG: Physical and immunochemical studies of the globulardomain of type IV collagen. Cryptic properties of the Goodpastureantigen. J Biol Chem 260:8564–8570, 1985

9. HUDSON BG, WIESLANDER J, WISDOM BJ JR, NOELKEN ME: Biologyof Disease. Goodpasture syndrome: Molecular architecture and func-tion of basement membrane antigen. Lab Invest 61:256–269, 1989

10. KALLURI R, GUNWAR S, REEDERS ST, MORRISON KC, MARIYAMA M,EBNER KE, NOELKEN ME, HUDSON BG: Goodpasture syndrome:Localization of the epitope for the autoantibodies to the carboxyl-terminal region of the a3(IV) chain of basement membrane collagen.J Biol Chem 266:24018–24024, 1991

11. KALLURI R, GATTONE VH, NOELKEN ME, HUDSON BG: The a3 chainof type IV collagen induces autoimmune Goodpasture syndrome. ProcNatl Acad Sci USA 91:6201–6205, 1994

Abbate et al: Intrathymic antigen injection and nephritis1560

12. KRONENBERG M: Self-tolerance and autoimmunity. Cell 65:537–542,1991

13. KAPPLER J, ROEHM N, MARRACK P: T cell tolerance by clonalelimination in the thymus. Cell 49:273–280, 1987

14. SPRENT J, LO D, GAO E-K, RON Y: T cell selection in the thymus.Immunol Rev 101:173–190, 1988

15. ELLISON GW, WAKSMAN BH: The role of the thymus in tolerance. IX.Inhibition of experimental autoallergic encephalomyelitis by intrathymicinjection of encephalitogen. J Immunol 105:322–326, 1970

16. GOSS JA, NAKAFUSA Y, ROLAND CR, HICKEY WF, FLYE WM:Immunological tolerance to a defined myelin basic protein antigenadministered intrathymically. J Immunol 153:3890–3898, 1994

17. KHOURY SJ, SAYEGH MH, HANCOCK WW, GALLON L, CARPENTERCB, WEINER HL: Acquired tolerance to experimental autoimmuneencephalomyelitis by intrathymic injection of myelin basic protein orits major encephalitogenic peptide. J Exp Med 178:559–566, 1993

18. KOEVARY SB, BLOMBERG M: Prevention of diabetes in BB/Wor ratsby intrathymic islet injection. J Clin Invest 89:512–516, 1992

19. POSSELT AM, BARKER CF, FRIEDMAN AL, NAJI A: Prevention ofautoimmune diabetes in the BB rat by intrathymic islet transplanta-tion at birth. Science 256:1321–1324, 1992

20. GERLING I, SERREZE D, CHRISTIANSON S, LEITER E: Intrathymic isletcell transplantation reduces b-cell autoimmunity and prevents diabe-tes in NOD/Lt mice. Diabetes 41:1672–1676, 1992

21. HEROLD KC, MONTAG AG, BUCKINGHAM F: Induction of tolerance toautoimmune diabetes with islet antigens. J Exp Med 176:1107–1114, 1992

22. KALLURI R, SUN MJ, HUDSON BG, NELLSON EG: Goodpastureautoantigen: Structural delineation of two immunologically privilegedepitopes on the a3 chain of type IV collagen. J Biol Chem 271:9062–9068, 1996

23. LANGEVELD JPM, WIESLANDER J, TIMONEDA J, MCKINNEY P, BUT-KOWSKI RJ, WISDOM BJ, HUDSON BG: Structural heterogeneity of thenoncollagenous domain of basement membrane collagen. J Biol Chem263:10481–10488, 1988

24. KALLURI R, WILSON CB, WEBER M, GUNWAR S, CHONKO AM,NEILSON EG, HUDSON BG: Identification of a3 chain of type IVcollagen as the common autoantigen in anti-GBM disease and Good-pasture syndrome. J Am Soc Nephrol 6:1178–1185, 1995

25. KALLURI R, MELENDEZ E, RUMPF WK, SATTLER K, MULLER GA,STRUTZ F, NEILSON EG: Specificity of circulating and tissue-boundautoantibodies in Goodpasture syndrome. Proc Assoc Am Phys 108:1–6, 1996

26. BUTKOWSKI RJ, LANGEVELD JPM, WIESLANDER J, HAMILTON J,HUDSON BG: Localization of the Goodpasture epitope to a novelchain of basement membrane collagen. J Biol Chem 262:7874–7877, 1987

27. READ SM, NORTHCOTE DH: Minimization of variation in the responseto different proteins of the Coomassie blue G-dye binding assay forprotein. Anal Biochem 116:53–64, 1981

28. WOODROFFE AJ, WILSON CB: An evaluation of elution techniques inthe study of immune complex glomerulonephritis. J Immunol 118:1788–1794, 1977

29. GUNWAR S, BALLESTER F, KALLURI R, TIMONEDA J, CHONKO CM,EDWARDS S, NOELKEN ME, HUDSON BG: Glomerular basementmembrane: Identification of the dimeric subunits of the non-collage-nous domain (Hexamer) of collagen IV and the Goodpasture antigen.J Biol Chem 266:15318–15324, 1991

30. NEILSON EG, KALLURI R, SUN MJ, GUNWAR S, DANOFF T, MARIYAMAM, MYERS JC, REEDERS ST, HUDSON BG: Specificity of Goodpastureautoantibodies for the recombinant noncollagenous domains of hu-man type IV collagen. J Biol Chem 268:8402–8405, 1993

31. LERNER RA, GLASSOCK RJ, DIXON FJ: The role of anti-glomerularbasement membrane antibody in the pathogenesis of human glomer-ulonephritis. J Exp Med 126:989–1004, 1967

32. KARNOVSKY MJ: Formaldehyde-glutaraldehyde fixative of high osmo-larity for use in electronmicroscopy. (abstract) J Cell Biol 27:137A,1965

33. WICK G, VON DER MARK H, DIETRICH H, TIMPL R: Globular domainof basement membrane collagen induces autoimmune pulmonarylesions in mice resembling human Goodpasture disease. Lab Invest55:308–317, 1986

34. BOLTON WK, LUO A-M, FOX P, MAY W, FOX J: Goodpasture’sepitope in development of experimental autoimmune glomerulone-phritis in rats. Kidney Int 49:327–334, 1996

35. SADO Y, BOUTAUD A, KAGAWA M, NAITO I, NINOMIYA Y, HUDSONBG: Induction of anti-GBM nephritis in rats by recombinanta3(IV)NC1 and a4(IV) NC1 of type IV collagen. Kidney Int 53:664–671, 1998

36. DOWNIE GH, ROHOLT OA, JENNINGS L, BLAU M, BRENTJENS JR,ANDRES GA: Experimental anti-alveolar basement membrane anti-body-mediated pneumonitis. II. Role of endothelial damage andrepair, induction of autologous phase, and kinetics of antibodydeposition in Lewis rats. J Immunol 129:2647–2652, 1982

37. JENNINGS L, ROHOLT OA, PRESSMAN D, BLAU M, ANDRES GA,BRENTJENS JR: Experimental anti-alveolar basement membrane anti-body-mediated pneumonitis. I. The role of increased permeability of thealveolar capillary wall induced by oxygen. J Immunol 127:129–134, 1981

38. SALANT DJ: Immunopathogenesis of crescentic glomerulonephritisand lung purpura. Kidney Int 32:408–425, 1987

39. ZIMMERMAN SW, GROEHLER K, BEIRNE G: Hydrocarbon exposureand chronic glomerulonephritis. Lancet 2:199–201, 1975

40. DONAGHY M, REES AJ: Cigarette smoking and lung haemorrhage inglomerulonephritis caused by autoantibodies to glomerular basementmembrane. Lancet 2:1390–1392, 1983

41. QUELUZ TH, PAWLOWSKI I, BRUNDA MJ, BRENTJENS JR, VLADUTIUAO, ANDRES G: Pathogenesis of an experimental model of Goodpas-ture’s hemorrhagic pneumonitis. J Clin Invest 85:1507–1515, 1990

42. REMUZZI G, NORIS M, BENIGNI A, IMBERTI O, SAYEGH MH, PERICON: Thromboxane A2 receptor blocking abrogates donor-specific un-responsiveness to renal allografts triggered by thymic recognition ofhistocompatibility allopeptides. J Exp Med 180:1967–1972, 1994

43. KALLURI R, DANOFF TM, OKADA H, NEILSON EG: Susceptibility toanti-glomerular basement membrane disease and Goodpasture syn-drome is linked to MHC class II genes and the emergence of Tcell-mediated immunity in mice. J Clin Invest 100:2263–2275, 1997

44. OHTSURU I, MATSUO H, FUKUDOME T, SUENAGA A, TSUJIHATA M,NAGATAKI S: ’Split tolerance’ induction by intrathymic injection ofacetylcholine receptor in a rat model of autoimmune myastheniagravis; implications for the design of specific immunotherapies. ClinExp Immunol 102:462–467, 1995

45. BONOMO A, MATZINGER P: Thymus epithelium induces tissue-specifictolerance. J Exp Med 177:1153–1164, 1993

46. RANSOM J, WU R, FISCHER M, ZLOTNIK A: Antigen presenting abilityof thymic macrophages and epithelial cells: Evidence for defects in theantigen processing function of thymic epithelial cells. Cell Immunol134:180–190, 1991

47. GAO E, LO D, SPRENT J: Strong T cell tolerance on parent F1 bonemarrow chimeras prepared with supraletal irradiation. J Exp Med171:1101–1121, 1990

48. SPEISER DE, LEES RK, HENGARTNER H, ZINKERNAGEL RM, MAC-DONALD HR: Positive and negative selection of T cell receptor Vbdomains controlled by distinct cell populations in the thymus. J ExpMed 170:2165–2170, 1989

49. OLUWOLE SF, JIN M-X, CHOWDHURY NC, OHAJEKWE OA: Effective-ness of intrathymic inoculation of soluble antigens in the induction ofspecific unresponsiveness to rat islet allografts without transientrecipient immunosuppression. Transplantation 58:1077–1081, 1994

50. BOREL Y, FAUCONNET M, MIESCHER PA: Selective suppression ofdelayed hypersensitivity by the induction of immunological tolerance.J Exp Med 123:585–598, 1966

51. CHILLER JM, HABICHT GS, WEIGLE WO: Kinetic differences inunresponsiveness of thymus and bone marrow cells. Science 171:813–815, 1971

52. RAJEWSKY K, BRENIG C: Paralysis to serum albumins in T and Blymphocytes in mice: Dose dependence, specificity and kinetics ofescape. Eur J Immunol 4:120–125, 1974

53. BURNSTEIN HJ, SHEA CM, ABBAS AK: Aqueous antigens induce invivo tolerance selectivity in IL-2 and IFN-gamma-producing (Th1)cells. J Immunol 148:3687–3691, 1992

54. KAGNOFF MF: Effects of antigen-feeding on intestinal and systemicimmune responses II. Suppression of delayed-type hypersensitivityreactions. J Immunol 120:1509–1513, 1978

55. CHALLACOMBE SJ, TOMASI TB: Systemic tolerance and secretory immu-nity after oral immunization. J Exp Med 153:1459–1472, 1980

56. HUSBY S, MESTECKY J, MOLDOVEANU Z, HOLLAND S, ELSON CO: Oraltolerance in humans. T cell but not B cell tolerance after antigenfeeding. J Immunol 152:4663–4670, 1994

Abbate et al: Intrathymic antigen injection and nephritis 1561